Agressores e Alvos: O mapa dos ataques identitários no Reddit brasileiro

• Beatriz Reis Gama Barbosa UFMG
• Izadora Monken Ganem UFMG
• João Lucas Simões Moreira UFMG
• Ana Paula Couto da Silva UFMG

DOI: https://doi.org/10.5753/webmedia.2026.30955

Resumo

Social media are virtual environments in which interactions are not always positive, especially since user identification is often not mandatory. Among different types of toxic content, attacks targeted at identity groups are particularly deserving of attention due to their potential to reinforce biases and marginalize already vulnerable communities. In this context, this work introduces a methodology to identify toxic content targeted at different identity groups in online platforms. This study identified approximately 73.4 thousand hostile messages on Brazilian Reddit, which allowed us to characterize the main groups targeted by the attacks. The analyses revealed that these attacks focused on subreddits with polarizing themes, such as politics and football, and the topic modeling highlighted different patterns between the attack types, with a strong recurrence of themes related to gender, mental health and political figures. Furthermore, the study of structural abuse patterns revealed an elevated prevalence of spurious abuse, which means most attacks received by a target occur in an isolated interaction with the aggressor. Our results suggest that specific and localized intervention measures are promising strategies to mitigate this behavior on social media and can support future user-targeted awareness campaigns, as well as aid the development of automatic moderation algorithms.

Palavras-chave: Toxicidade Online, Discurso de Ódio, Anotações Semânticas, Processamento de Linguagem Natural, Modelagem de Tópicos

Referências

Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 4171–4186.

Arne Dreißigacker, Philipp M"uller, Anna Isenhardt, and Jonas Schemmel. 2024. Online Hate Speech Victimization: Consequences for Victims’ Feelings of Insecurity. Crime Science 13 (2024), 4. DOI: 10.1186/s40163-024-00204-y

Paula Fortuna and S’ergio Nunes. 2018. A Survey on Automatic Detection of Hate Speech in Text. Comput. Surveys 51, 4 (2018), 85:1–85:30. DOI: 10.1145/3232676

P. T. M. Gibrim, M. V. G. Ribeiro, L. S. Moreira, P. de F. Melo, and J. C. S. Reis. 2025. I’m not Welcome Here – Uma Revisão de Escopo sobre Toxicidade e Discurso de Ódio em Jogos Digitais. In Anais do Simpósio Brasileiro de Sistemas Multimídia e Web (WebMedia). [link]

Maarten Grootendorst. 2020. BERTopic: Leveraging BERT and c-TF-IDF to Create Easily Interpretable Topics. DOI: 10.5281/zenodo.4430182

Deepak Kumar, Jeffrey Hancock, Kurt Thomas, and Zakir Durumeric. 2023. Understanding the Behaviors of Toxic Accounts on Reddit. In Proceedings of the ACM Web Conference 2023 (WWW ’23). DOI: 10.1145/3543507.3583522

Giovana Piorino, Vitor Moreira, Luiz Henrique Quevedo Lima, Adriana Silvina Pagano, and Ana Paula Couto da Silva. 2024. Análise de sentimentos de conteúdo compartilhado em comunidades brasileiras do Reddit: Avaliação de um conjunto de dados rotulados por humanos. In Proceedings of the 30th Brazilian Symposium on Multimedia and the Web (WebMedia). 54–62. DOI: 10.5753/webmedia.2024.242020

Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing. 3982–3992.

Punyajoy Saha, Kiran Garimella, Narla Komal Kalyan, Saurabh Kumar Pandey, Pauras Mangesh Meher, Binny Mathew, and Animesh Mukherjee. 2023. On the Rise of Fear Speech in Online Social Media. Proceedings of the National Academy of Sciences 120, 11 (2023), e2212270120. DOI: 10.1073/pnas.2212270120

Wenhui Wang, Furu Wei, Li Dong, Hangbo Bao Bao, Nan Yang, and Ming Zhou. 2020. MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers. In Advances in Neural Information Processing Systems, Vol. 33. DOI: 10.48550/arXiv.2002.10957