Bengali and English Languages Fake-news Identification

Artificial Intelligence (AI) is revolutionizing various fields by automating tasks that typically require human intelligence, creating new possibilities for improving service delivery, ensuring policy compliance, and detecting fraud. However, one of AI’s most crucial applications lies in combating the spread of fake news. The proliferation of false information online poses significant threats to informed decision-making and public opinion. This research aims to improve fake news detection for both English and Bengali news articles. By addressing the data imbalance, especially in Bengali fake news, we curated a more balanced dataset. Utilizing transformer-based pre-trained models, we achieved notable enhancements in identifying fake news and applied interpretable models to clarify the rationale behind these classifications. Our work offers a novel dataset for fake news in the Bengali language and presents a system that not only determines whether a Bengali or English language news is either fake or real but also highlights the specific words influencing the prediction, making the decision process more interpretable for users.