Published Paper Details:

With the rise of deepfake content on social media, distinguishing between genuine and AI generated tweets has become a major challenge. This project proposes an advanced detection framework that utilizes Fast-Text embeddings and deep learning models to identify manipulated text. Fast-Text is chosen for its ability to capture semantic meaning, sub word information, and contextual nuances in social media text, including slang and misspellings. The system integrates LSTM, GRU, and Transformer models to enhance classification accuracy, following a structured workflow of data collection, text preprocessing, Fast Text-based feature extraction, and model training. To ensure reliable detection, the system is evaluated using key metrics like accuracy, precision, recall, and F1 score, demonstrating superior performance over traditional methods. By effectively handling out-of-vocabulary (OOV) words and noisy tweet data, the proposed framework provides a scalable and robust solution for detecting machine-generated tweets. Future enhancements may include real-time detection, integration with BERT, RoBERT, or GPT