Last fall, we featured an extensive interview with Petter Törnberg of the University of Amsterdam, who studies the underlying mechanisms of social media that fuel its most harmful effects: partisan echo chambers, the concentration of influence among a small elite (attention inequality), and the amplification of extreme, divisive voices. His outlook on social media’s future was pessimistic.
Törnberg’s research concluded that while numerous platform-level interventions have been proposed to address these issues, none are likely to succeed. The problem isn’t algorithms, non-chronological feeds, or human tendencies toward negativity. Instead, the dynamics driving these negative outcomes are structurally embedded in social media’s very architecture. Without a fundamental redesign that disrupts these dynamics, we’re likely trapped in endless toxic feedback loops.
New Research Reinforces Structural Flaws in Social Media
Since that interview, Törnberg has published two new papers and one preprint, expanding on the idea that social media operates differently from the physical world, with unintended consequences. His first new paper, published in PLoS ONE, specifically examines the echo chamber effect using a combined approach of standard agent-based modeling and large language models (LLMs). This method essentially creates AI personas to simulate online social media behavior.
Why Interventions Fail: The Core Problem
Törnberg’s work highlights that social media’s structural flaws are not superficial—they are fundamental to how these platforms function. Echo chambers, attention inequality, and the amplification of divisive content are not bugs but features of the system. Current interventions, such as algorithmic adjustments or content moderation policies, only address symptoms rather than the root cause.
What’s Next? The Search for a Fundamental Redesign
Törnberg’s research suggests that meaningful change will require rethinking social media’s architecture from the ground up. This could involve redesigning how information spreads, how influence is distributed, or how user interactions are structured. Without such a transformation, toxic feedback loops will likely persist, perpetuating the same harmful dynamics we see today.