started · updated
Researchers develop virtual twin to study social media echo chambers
A research team led by the National Research Council’s Institute of Information Science and Technologies (CNR-ISTI) has developed YSocial, a digital platform designed to simulate social network dynamics. The system acts as a virtual twin, using thousands of artificial users generated by Large Language Models (LLMs) to replicate how algorithms and user behaviors interact.
By creating agents with specific ages, interests, and political opinions, researchers can observe the development of political polarization and the formation of echo chambers in a controlled environment. This approach addresses the difficulty of studying real-world platforms, where data access is limited and algorithmic transparency is low. The study, published in the journal Big Data & Society, aims to provide insights into how recommendation systems can limit voter access to diverse viewpoints, particularly during significant events like the United States midterm elections.
Entities
CNR-ISTI · Claude Bernard University Lyon 1 · National Research Council · University of Pisa · University of Trento