A study shows: Social media is probably beyond saving

[Christian]

For years, platforms, governments, and researchers have been trying to get a handle on the negative effects of social media. But filter bubbles, toxic debates, polarization, and radicalization are not as easy to curb as some might think.
A recent study by Petter Törnberg and Maik Larooij from the University of Amsterdam suggests that the problem not only defies a quick solution, but may even be structurally unsolvable.
In their analysis, the researchers combined agent-based modeling with large language models (LLMs) to create realistic simulations of social networks. This involves simulating many virtual individuals with different opinions and behaviors who interact online. The aim is to observe how typical dynamics such as echo chambers or polarization arise from their interactions. The results were sobering.
Without any algorithmic reinforcement, complex rules, or “malicious” platform designs, destructive dynamics and echo chambers developed, along with extreme inequality in the distribution of attention and the dominance of provocative content. Even the simplest network of posting, reposting, and following is enough to create a toxic system, the researcher explains in an interview with Ars Technica.