The rapid proliferation of misinformation across social media platforms presents significant challenges for governmental institutions in maintaining effective fact-checking processes. This paper introduces DemocraticNet, an automated multi-class multi-label classification pipeline that leverages state-of-the-art transformer architectures to assist in news classification and verification workflows. Our system employs an ensemble of pre-trained language models, including BERT, DeBERTaV3, RoBERTa, and XLNet, fine-tuned on a curated dataset of news articles across various ministerial domains. The pipeline performs two primary tasks: (1) multi-label classification to identify relevant government ministries for each news item, achieving a top-3 accuracy of 85.043% and cross-entropy loss of 6.063%. At the time of writing, no dataset or literature exists on multi-class news categorization, and no study has addressed multi-class classification of news, making this work the first to pioneer it. The experimental results demonstrate the model’s capability to effectively categorize news items across multiple ministerial jurisdictions while providing initial credibility indicators. This automated classification approach could potentially support existing manual verification workflows by providing systematic initial assessments. We also discuss the system’s current limitations and important considerations for deploying such automated news classification systems in democratic contexts.