Name: ATHUS ASSUNÇÃO CAVALINI
Publication date: 28/11/2024
Examining board:
Name |
Role |
|---|---|
| ANDRE GEORGHTON CARDOSO PACHECO | Examinador Interno |
| FÁBIO LUIZ MALINI DE LIMA | Examinador Externo |
| GIOVANNI VENTORIM COMARELA | Presidente |
| JUSSARA MARQUES DE ALMEIDA GONÇALVES | Examinador Externo |
| MAGNOS MARTINELLO | Examinador Interno |
Summary: The popularization of social networks and their evolution into digital platforms have
triggered a profound revolution in human communication and interaction processes. While
this advancement offers countless opportunities, it also brings complex challenges, such as
informational disorder. Social platforms provide a structure that enables the rapid spread
of content, including misleading or harmful information, with a high potential for virality.
This phenomenon becomes especially critical in public health contexts, as seen during
the COVID-19 pandemic, when the spread of misinformation led to increased vaccine
hesitancy, compromised health campaigns, and exacerbated the pandemic’s effects. In
this context, this study investigates the spread of misinformation within the anti-vaccine
community on Telegram, one of the world’s most popular messaging platforms, using data
science and social computing methods.
Through the analysis of nearly 10 million messages shared across 779 channels and groups,
this research aimed to understand the dynamics of this community, including its structure,
information flow, and user engagement. Data were collected using a sampling methodology
for partially observable networks, enabling an effective mapping of the network’s structure.
To detect misinformation, a machine learning model was developed specifically to categorize
anti-vaccine messages in terms of their veracity and potential for harm. Additionally, the
level of content toxicity was also analyzed.
The results indicated that the anti-vaccine community on Telegram has a high level of
engagement and that misinformation narratives follow specific patterns. The analyses of
misinformation, toxicity, and engagement revealed correlations between the type of content
and its impact on the audience, providing valuable insights into possible misinformation
strategies within this context.
This work, therefore, contributes to a deeper understanding of the phenomenon of informa-
tional disorder on social platforms, evaluating methodologies, tools, and metrics suitable
to the study’s objectives and contributing to the development of effective methodologies
and tools to combat misinformation in specific contexts.
