Date
Share

Emanuele Brugnoli

Sony CSL – Rome

I’ve always been fascinated by the tools mathematics offers to describe and understand complex problems in an extraordinarily simple and elegant way. This fascination led me to pursue a PhD in Mathematics and Computer Science (University of Perugia, 2016), where I focused on closed-form solutions to enumeration problems on graphs. Since then, my research journey has taken me through several leading institutions in Italy, including the University of Palermo, the Institute for Complex Systems (ISC-CNR), the Italian Communications Authority (AGCOM), and the Enrico Fermi Research Centre (CREF).
In these roles, I explored the use of graphs as geometric structures for representing large-scale data, particularly in the context of information flows, opinion dynamics, and the emergence of bias—especially within online environments. These interdisciplinary experiences have deepened my interest in complexity science and its applications across media and society.
I’m currently part of the Infosphere team at Sony CSL in Rome, where I integrate natural language processing and statistical learning to develop models and tools that help us better understand online discourse and address the distortions that can undermine it. My aim is to contribute to a more transparent and navigable information ecosystem.

Opinion Formation: Biases in News Reporting and Consumption

News media are inherently selective—both in what they report (selection bias) and how they present it (narrative bias). Selection bias refers to which events are covered or ignored, while narrative bias concerns how stories are framed through language, tone, and emphasis, shaping public perception. These production-side biases interact with confirmation bias on the consumer side: individuals tend to seek out news that aligns with their preexisting beliefs. This mutual reinforcement creates a feedback loop in which media outlets increasingly tailor content to ideological audiences—amplifying polarization, echo chambers, and misinformation. This seminar explores these dynamics through the case of the vaccine debate in Italy, offering a quantitative analysis of both selection and narrative biases. Results show that negative and extreme content attracts greater attention, and ideologically similar outlets tend to share their audiences—further reinforcing echo chamber effects. While narrative bias is effectively captured by third-party assessments, selection bias remains more elusive, highlighting the inadequacy of current evaluation criteria. Understanding and quantifying the different forms of bias is essential for promoting a more accurate and inclusive representation of global events in online discourse. Upholding journalistic integrity requires deliberate efforts to mitigate bias, ensure transparency, and embrace diversity in reporting practices and perspectives.