Multilingual Fake News and Hate Speech Detection System

This paper presents the development of a multilingual fake news and hate speech detection system using transformer-based models. The system is designed to address the increasing spread of misinformation and harmful content across digital platforms, especially in multilingual and code-mixed environments such as Tamil–English (Tanglish). By leveraging advanced transformer architecture like XLM-RoBERTa, the system effectively captures contextual and semantic relationships in text, enabling accurate classification of content into categories such as real or fake news and neutral, offensive, or hate speech. The model is trained and evaluated using real-world datasets including LIAR and HASOC, and the results demonstrate strong performance in terms of accuracy, precision, recall, and F1-score.   The findings of this study highlight the importance of using advanced natural language processing techniques for content moderation in modern digital platforms. The proposed system provides a scalable and efficient solution that can be integrated into real-time applications such as social media monitoring, fact-checking systems, and browser extensions. By enabling automatic detection of fake news and hate speech, the system helps in reducing the spread of misinformation and promoting safer online communication. In conclusion, this work contributes towards the development of intelligent and reliable content moderation systems. Future work can focus on improving model performance, supporting additional languages, enhancing explainability, and extending the system to handle multimodal data such as images and videos for more comprehensive analysis.

1)      Ahuja, N., & Kumar, S. (2023). Mul-FaD: Attention-based multilingual fake news detection using deep learning techniques. In Proceedings of the International Conference on Artificial Intelligence and Data Science. 2)      Ghosh, K., Singh, A., & Sharma, P. (2025). Hate speech detection in low-resourced Indian languages using transformer-based models. In Proceedings of the International Conference on Computational Linguistics. 3)      Kumar, R., Verma, S., & Patel, M. (2025). Explainable multilingual fake news detection using XLM-RoBERTa with LIME and SHAP techniques. In Journal of Artificial Intelligence Research. 4)      Das, A., Roy, B., & Chakraborty, S. (2024). Multi-task learning approach for hate speech detection using XLM-RoBERTa. In Neurocomputing, Elsevier. 5)      Yadav, A., Mishra, R., & Gupta, V. (2024). Code-mixed Hinglish hate speech detection using transformer-based models. In Proceedings of COLING. 6)      Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of NAACL-HLT. 7)      Conneau, A., Khandelwal, K., Goyal, N., Chaudhary, V., Wenzek, G., Guzmán, F., et al. (2020). Unsupervised cross-lingual representation learning at scale. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (pp. 8440–8451). 8)      Ranasinghe, T., Orasan, C., & Mitkov, R. (2022). Multilingual hate speech detection using transformer-based models. In Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP). 9)      Alam, F., Cresci, S., Chakraborty, T., et al. (2023). Multilingual fake news detection using transformer-based models. In IEEE Access. 10)    Singhal, S., et al. (2020). SpotFake: A multi-modal framework for fake news detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).